With gasoline prices on the rise again, so with it are consumers unhappy about new-car fuel efficiency ratings. Among the latest, nearly 1 million Hyundai and Kia owners appear to have been misled. The companies recently admitted that their claims of a 40-mpg fleet were a bit too optimistic. Based on an Environmental Protection Agency investigation resulting from complaints of overstated fuel-economy estimates, Hyundai and Kia have agreed to lower the posted mileage ratings on the majority of their 2012 and 2013 offerings. The downward revisions amount to 1 to 2 mpg on most models, with the largest correction being a 6-mpg reduction in the Kia Soul’s highway rating. Both companies plan to reimburse vehicle owners to cover additional fuel costs over what was originally promised.

In another case, based on claims from Consumer Reports and others, the EPA reportedly plans to investigate whether the 2013 Ford C-Max Hybrid really deserves its 47/47 city/highway ratings. The publication said in its tests, the car averaged between 33 and 39 mpg. It also notes that the 2013 Ford Fusion Hybrid uses the same powertrain and is also rated at 47/47 mpg, but only averaged 40 mpg in its testing. And last March, the EPA had BMW lower ratings for its 328i automatic by 1 mpg in city and 3 mpg on the highway.

How does this happen? Most of us assume the EPA tests and rates all new cars. In reality, the agency lets auto companies test their own cars and trucks according to government guidelines, and accepts the fuel-efficiency numbers the companies submit.

The agency only tests 10 to 15% of new models. Of those, most are selected at random and the rest usually involve new manufacturers, new technologies, or extremely high or low-rated vehicles. (To that list, we suggest the EPA add all Hyundais, Kias, and hybrids for the foreseeable future.)

Can manufacturers honestly “game” the system? Most experts say no. Hyundai and Kia, for instance, say their misstatements were due to “errors” at the automakers’ joint testing operations in Korea. In the lab, the vehicle sits on a dynamometer and is put through a series of tests to simulate “typical” city and highway driving. It is also run at high speeds, in cold temperatures, and with the A/C on. Wheel speeds are specified, as are the distances and number of stops. The detailed procedures are said to account for real-world factors such as aerodynamic drag and inertia, and final city and highway ratings are calculated from fuel-economy results from each of the five tests.

However, one problem is human drivers behind the wheel of the test cars control throttle and braking. Because they can’t exactly follow test requirements for speed and acceleration, there’s a permissible tolerance band. It’s not unreasonable to assume manufacturers know how to maximize ratings while staying within the limits of the protocol.

Even if testing is by the book, why doesn’t your mileage match the window sticker? In short, no one drives exactly like the EPA test and many everyday factors hurt mileage. Edmonds.com says heavy but underpowered vehicles match up worst.

In the end, when the showroom salesperson touts high mileage, it’s buyer beware. In a recent Gallup poll on trustworthy professions, car salespeople came in dead last — behind even Members of Congress.

© 2013 Penton Media, Inc.